key: cord-0252704-ie5ox69g authors: nan title: Translational diffusion of class II major histocompatibility complex molecules is constrained by their cytoplasmic domains date: 1989-12-01 journal: J Cell Biol DOI: nan sha: 44ca0c52f79a893a551da92dd78e599fbc74b060 doc_id: 252704 cord_uid: ie5ox69g Site-directed mutagenesis in vitro was used to introduce stop codons in the genomic DNA of the alpha and beta chains of the murine class II major histocompatibility complex antigen, I-Ak. Mutated DNA was transfected into B lymphoma cells that were then selected by neomycin resistance and for their ability to express I-Ak molecules on their plasma membrane. The translational diffusion coefficient (Dlat) of I-Ak molecules composed of a wild-type beta chain paired with an alpha chain missing either 6 or 12 amino acids from the cytoplasmic domain is on the average threefold higher than the Dlat of wild-type I-Ak molecules as measured by fluorescence photobleaching and recovery. The removal of 12 amino acids from the cytoplasmic domain of the beta chain did not change the Dlat value from that of wild-type I-Ak if the truncated beta chain was paired with a wild-type alpha chain. Removing all amino acids of the cytoplasmic domains of both the alpha and beta chains resulted in a 10-fold increase in the Dlat, the highest value for any of the truncated I-Ak molecules tested. These data indicate that the carboxy- terminal six amino acids of the cytoplasmic domain of the alpha chain and the six plasma membrane-proximal amino acids of the beta chain are important in constraining the translational diffusion of I-Ak molecules in the plasma membrane. T HE cytoplasmic domains of a number of transmembrane membrane proteins are thought to be important in the membrane biology of those proteins (3, 13, (16) (17) (18) 31) . Plasma membrane proteins with truncated cytoplasmic domains may not be correctly sorted or internalized and may lose signaling functions (22) (23) (24) 26) . Surprisingly, the translational diffusion coefficient (D,.t)' of proteins with such truncations is not found to differ from that of full-length molecules. Dj., of the EGF receptor (16) , class I major histocompatibility complex (MHC) antigen (10) , or G protein of vesicular stomatitis virus (25) is not changed when the cytoplasmic domains of these molecules are truncated by directed mutation of their genes. At the extreme, this involves a reduction from 542 to 9 residues in the EGF receptor (16) . These studies imply that the amino acids removed are not associated with cytoplasmic proteins in ways that constrain their translational diffusion. There is an apparent contradiction between the results on the cell physiology of truncated molecules and the measurements of their translational diffusion. Results of endocytosis experiments indicate that the molecules are coupled to the cell cytoskeleton and that this coupling is lost with truncation since internalization is affected by truncation. On the other hand, the results of translational diffusion experiments suggest that even wild-type, 1. Abbreviations used in this paper: D0~t, translational diffusion coefficient; FPR, fluorescence photobleachin 8 and recovery; MCF, mean channel fluorescence; MHC, major histocompatibility complex; PKC, protein kinase C. full-length molecules are not directly linked to the cytoskeleton since truncation of their cytoplasmic domains has no effect on D,.,. This contradiction may arise largely because translational diffusion and endocytosis of truncated molecules have been studied in different cell types. D~ of the truncated class I MHC mutants has been measured in L cells, in which class I molecules are not rapidly internalized. The effect of truncation on endocytosis of the same mutants is readily shown when they are expressed in T lymphoma cells, a cell type in which there is continuous cycling of class I MHC antigens between the surface and cytoplasm (4) . Similarly, D,,, of truncated EGF receptor was measured in COS cells, while effects of truncation on endocytosis of EGF receptor were shown in 3T3 cells (2) . Here we present the first results in which D,a, of truncated transmembrane molecules is measured in cells in which the wild-type molecules are functional. Truncation of the cytoplasmic domain affects D,a, in these cells; in the extreme truncation mutant, the D,,, is an order of magnitude larger than that of the wild-type molecule. Class II MHC molecules, Ia antigens in the mouse, are heterodimeric transmembrane proteins expressed on B lymphocytes, macrophages, and other antigen-presenting cells. There are two isotypes of Ia antigens in the mouse, I-A and I-E; each consists of a 34-36-kD ct chain and a 27-29-kD/~ chain. Each chain is composed of two external domains (ot~/ot2 or/3,/~2) and a connecting peptide that joins the ex-ternal domains to the transmembmne and cytoplasmic domains. There are two asparagine-linked oligosaccharide units on each c+ chain and one on the B chain. One function of Ia molecules is to present foreign peptide ~wt/[~wt antigens to T lymphocytes. Antigen presentation in B lymphocytes involves facilitated internalization of native antigen by antigen-specific surface Ig (signaling aspects reviewed in reference 6) followed by processing of the foreign antigen by denaturation and/or fragmentation. Ia molecules also function as receptors that influence the differentiative or proliferative events in B lymphocytes (6, 7) . Potentially, Ia molecules are coupled to the molecules of the signal-transducing cascade through the highly conserved amino acids of the transmembrane domain or cytoplasmic domain (12) . Ia molecules and surface Ig (which also provides signals to B cells) are similar in their ability to associate with insoluble cytoskeletal elements after receptor cross-linking and to remain associated with the cytoskeleton during solubilization of the cells (30) . These cytoskeletal associations may be important for the signaling function which Ia and Ig molecules provide. B cell lymphomas transfected with class II genes have also been used to study one aspect of B lymphocyte signaling, the translocation of protein kinase C (PKC) from the cytoplasm to the nucleus, that is induced by perturbation of membrane Ia molecules by Ia-specific rnAbs. This binding of surface Ia by mAbs presumably mimics the interaction of T cell receptors with peptide antigen-bearing Ia molecules. I-A k molecules consisting of truncated/~ chains paired with either fulllength ot chains (ctwt//~CT12) or with truncated ot chains that had no cytoplasmic domain amino acids (otCT12/flCT12) had altered PKC signaling phenotypes compared with wild-type I-A k molecules. Further, I-A k molecules that lacked cytoplasmic domains of both a and/3 chains (aCT12//3CT18) were unable to signal at all as measured by the translocation of PKC from the cytoplasm to the nucleus (28) . These results collectively suggest that the cytoplasmic domain of I-A k is functionally coupled to a signal transduction pathway and to the cytoskeleton in B lymphocytes. Given this coupling, it might be expected that deletions of all or part of Cells the cytoplasmic domain would affect the D,= of the mutant Ia molecules. Using fluorescence photobleaching and recovery (FPR) measurements, we have determined the D~,, of wild-type I-A k molecules and of mutant I-A k molecules lacking all or part of the cytoplasmic domain of one or both chains. Here we show that the D~= of I-A k molecule is a function of the length of its cytoplasmic domains. The pSV2-Neo construct (27) was a gift from Dr. Ed Palmer (National Jewish Center for Immunology and Respiratory Medicine, Denver, CO). The origin and generation of the a and fl I-A k genomic DNA constructs have been described (28) . In brief, site-directed mutagenesis in vitro was used to introduce double-stop codons that v~uld result in truncated a or fl chains lacking 6 and 12 or 12 and 18 amino acids, respectively. Uamb+nne I RHRsQK Transfectants were generated by introducing pSV2-Neo DNA and ~ and genomic DNA into the Ia-negative B cell iymphoma M12.C3 by electroporation. The transfecled cells were selected and maintained using (3418 (neomycin) at a final concentration of 300 t~g/ml. Transfectants expressing memk k brahe I-A were chosen based on flow cytometry using anti-l-A -specific mAb (Fig. 2) . mAbs were purified from tissue culture supernatants by affinity chromatogk raphy on protein A-Sepharose. The anti-l-A mAb, 39], is specific for the c~ chain (21) . The anti-I-A k mAb, 10-3.6, is specific for the fl chain (19) . The anti-L d mAb, 28-14-8, was used as a control to assess for nonspecific effect on plasma membrane protein Dl,, values (20),2-8 mg of mAb was digested with 1:100 (wt/wt) papain (Sigma Chemical Co., St. Louis, MO) at pH 5.2 for 10 min. The digest was fractionated on Sephadex G-100, and the first included peak was refractionated on protein A-sepharose to remove traces of IgG. The purified Fab was concentrated to 5-10 mg/mi protein, dialyzed into 0.2 M Na2CO3 buffer, pH 9.5, and reacted with 15 #g FITC/ mg protein at 0oc for I h. The reaction mixture was purified by passage c~er Sephadex G-25. All conjugates had 3-4 tool fluoresccin coupled/mol Fab. In an FPR experiment, diffusion in the plane of the membrane, translational diffusion, is measured in terms of recovery of fluorescence after partly bleaching a small area on an otherwise uniformly labeled cell surface (8, I I I II IIII I I I I 14). The recovery curve contains information about the Dlat of the mAbbound molecules and about the fraction, R, of the population of mAb-bound molecules that are free to diffuse, on a time scale of minutes, in the experiment. Dl,t of a given molecule is a function of the radius of the diffusing species. Hence, monovalent labels-i.e., Fab fragments (which do not cross-link bound molecules)-give more reliable estimates of Dlat than do larger and bivalent IgG labels. The mobile fraction, R, is influenced by the intensity and specificity of binding of the Fab, as well as by the biology of the ceils. Calculations and control experiments indicate that the measuring and bleaching beams do not cause detectable cross-linking of, or damage to, the low concentrations of molecules of interest in an FPR experiment (9) . In our experiments, cells were labeled with FITC-Fab for 15-30 min and washed four times in Hepes-buffered Hanks' balanced salt solution. Most measurements were made on cells in suspension and at 20"C. Our FPR machine has been described elsewhere (8) . In one experiment, there was no difference between the Dl=t measured on cells in suspension and Dtat measured on cells adherent to coverslips (data not shown). There was also no difference between Dlat at 20°C and Dlat at 30°C (data not shown). In experiments with MI2.C3-10-D2, the transfectant that expresses the lowest level of I-A k, recovery of fluorescence was reduced if cells were incubated with goat anti-mouse IgG before measurement (data not shown). This demonstrates that the fluorescence measured is from mAb bound to surface I-A k and is not cell autofluorescence. The values of Dlat typically are not a normal distribution. Accordingly, geometric mean Dlat were calculated from normally distributed laD. Signifieances of differences between mean InDRit were compared in Student's t test. p < 0.05 was considered significant. Transfectants expressing various combinations of truncated t~ or B chains were generated, and the translational diffusion of the class II molecules was measured by FPR. Fig. 2 shows representative fluorescence profiles of the transfectants expressing either wild-type or mutant I-A ~ molecules. The level of I-A k expressed at the cell surface differs by no more than twofold between these transfectants, with the exception of the otCT12//~CT18 truncation phenotype transfectant. It can be seen from Table I that removal of the carboxy-terminal 12 (ofa total of 18) amino acids of the/3 chain does not affect D~,~. Wild-type I-A k and o~wt//~CTl2 I-A ~ molecules have the same Dt~. This contrasts with the effect measured after the loss of as few as six amino acids from the tx chain. D~, of I-A k molecules of the ctCT6 or the t~CT12 phenotypes paired with a wild-type B chain is significantly higher than D~, of wild-type I-A k molecules. It seems that the ot chain of the I-A k molecule constrains its diffusion in the plasma membrane. This conclusion is supported by the results with the transfectant M12.C3-5-B2 with the I-A ~ truncation phenotype otCT12//~CTI2. D~, of otCTI2//~CTI2 I-A ~ molecules does not differ from that of transfectants of the otCT6/Bwt and o~CT12/Bwt phenotypes. It appears from these results that 12 of the 18 amino acids of the cytoplasmic domain of the/~ chain have little influence on the Dn~ of I-A k molecules and that the tx chain is primarily responsible for constraining the translational diffusion of I-A k. To assess the contribution of the six membranepromixal amino acids of the cytoplasmic domain of the/3 chain, we generated transfectants expressing I- For ease of comparison, the normalized fold increase of D~,, for individual I-A k phenotypes is shown in Fig. 3 . The order of increasing D~t is otwt/Bwt = txwt/BCT12 < txCT6/ /~wt = aCT12//~wt = oA2T12/BCTI2 < CT12/CTI8. The fold increase from wild type describing this order is 1 = 1.5 < 3.1 = 3.0 = 3.5 < 9.9. Statistical significance for various comparisons of the D~,t is summarized in the legend for Table I. All comparisons between D~, of I-A k on individual Table I. transfectants are only made within an experiment. Dj,, were independent of the differences between experiments in the mobile fractions (see below). The mobile fraction of wild-type and mutant I-A k antigens was constant within an experiment with one exception. The mobile fraction of the tailless I-A k molecule (t~CT12//~CT18) was significantly (p < 0.05) higher than that of the wild-type I-A k in both experiments 3 and 4. To ensure that D~,, measured for the transfectant M12.C3-10-D2 (otCTI2//3CT18) was in fact due to the truncation phenotype of the I-A k molecule and not the result of the low concentration of I-A k protein, we used cell sorting to enrich for M12.C3-10-D2 cells that expressed the highest level of the t~CT12/BCT18 I-A k molecule. A fluoresceinated ct chainspecific mAb was used to label cells that were sorted on a cell sorter (Epics 751; Coulter Electronics Inc., Hialeah, FL), and the 10% of the cells expressing the highest levels of trun-cated I-A k were selected. The resulting sorted population is shown in Fig. 2 D. A normally distributed population of ¢xCT12//3CT18 I-A k expressing cells can be seen using either directly conjugated Fab of mAb 39J (